Ultralocalized thermal reactions in subnanoliter droplets-in-air

Abstract

Miniaturized laboratory-on-chip systems promise rapid, sensitive, and multiplexed detection of biological samples for medical diagnostics, drug discovery, and high-throughput screening. Within miniaturized laboratory-on-chips, static and dynamic droplets of fluids in different immiscible media have been used as individual vessels to perform biochemical reactions and confine the products. Approaches to perform localized heating of these individual subnanoliter droplets can allow for new applications that require parallel, time-, and space-multiplex reactions on a single integrated circuit. Our method positions droplets on an array of individual silicon microwave heaters on chip to precisely control the temperature of droplets-in-air, allowing us to perform biochemical reactions, including DNA melting and detection of single base mismatches. We also demonstrate that ssDNA probe molecules can be placed on heaters in solution, dried, and then rehydrated by ssDNA target molecules in droplets for hybridization and detection. This platform enables many applications in droplets including hybridization of low copy number DNA molecules, lysing of single cells, interrogation of ligand-receptor interactions, and rapid temperature cycling for amplification of DNA molecules.

Fig. 1.

Device and methodology schematic. (A) Cross-section of device with a droplet is shown. The left side shows an unheated droplet with the DNA FRET construct in the double-stranded form. The right side shows a heated droplet where the FRET construct has denatured, resulting in an increase in fluorescence. (B) A microcapillary pressure injection system is used to spot droplets on individual devices. An AC signal between the shorted source drain and the back gate of a device is used to heat the droplet. (C) A top view of ∼225 pL droplet placed on a heating element. The heating element is 2 µm wide in a 20 µm × 20 µm release window. Scale bar, 100 µm. (D) An array of droplets is spotted on linked devices. Eleven linked on left module and 11 linked on right module. Scale bar, 100 µm.

Fig. 2.

Single droplet melting curves. (A) A melting curve from commercial real-time PCR machine shows an increase in fluorescence as the FRET construct denatures. (B) Derivative of A, the peak of which gives the melting temperature of the FRET construct shown in Table 1. (C) On-chip fluorescence data through a voltage sweep from 0 to 40V_rms. (D) Derivative of plot C showing the melting voltage of the constructs. Averages and SDs across multiple chips are shown in SI Appendix, Table S1. (E) Simulation versus experimental results for temperature-voltage calibration curve. Different fits are shown with R² values of linear, 0.854; cubic, 0.586; and theory’s fit, 0.935. (F) Simulation of the time it takes for the temperature to stabilize within the droplet.

Fig. 3.

Parallel droplet heating of multiple constructs. (A) A sequence of images showing the process of heating of linked devices for plots B and C. Each droplet contains a unique FRET construct with a different melting temperature (50, 61, and 80 °C). (B) A plot of the raw fluorescence data from the droplets during the voltage sweep. (C) The derivative of B provides the melting voltage for each of the constructs. SI Appendix,Table S3 provides averages and SDs for melting curves performed on multiple devices and chips. D and E provide a second example of linked device heating. In this example, it is possible to discern between two fully complementary strands and a heteroduplex that contains a single base mismatch. SI Appendix, Table S4 provides average and SDs for the melting voltage across multiple chips.

Fig. 4.

Probe DNA dehydration with target DNA rehydration. (A) An example of the process flow is presented. A, i shows the devices before dehydration of the probe ssDNA. After spotting of the probe ssDNA (A, ii), the DNA in solution is allowed to dehydrate, leaving behind residual salts and DNA. A, iii shows a fluorescent image of the dried ssDNA spot. The fluorescence intensity is high without the presence of the FRET quencher. A, iv shows the rehydration of devices 1, 2, and 3. The initial fluorescence before denaturation is show in A, v. The fluorescence intensity is lower than A, iii due to the introduction of the FRET quencher and the DNA hybridization. (B and C) A melting curve of three spots that have been rehydrated with a complementary target sequence. The increase in fluorescence shows a distinct, single peak. (E) This implies that the DNA has hybridized properly without unwanted heterodimer or self-dimer formation. (D and E) A test for specificity in the process. Spots of a dried probe sequence were rehydrated with a complementary sequence, a noncomplementary sequence, or water. A distinct peak in the derivate in E implies a matching sequence.